Darren,
I actually feel better knowing that there wasn't a real easy solution
:-) I did some experimenting and came up with a way that seems
relatively accurate for what I want.
>>
cd /usr/openv/netbackup/db/images/
if [ -f /tmp/stats ]
then
rm /tmp/stats
fi
ls > /tmp/clients
for CLIENT in `cat /tmp/clients`
do
cat $CLIENT/STREAMS |grep -v ^T | awk '{print $2,$4,$8}' |grep -v ^0
|sed 's/$/ '$CLIENT'/' >> /tmp/stats
done
<<
This gives me output looking like this for /stand on the host "silo-1":
1171155594 cumulative-inc /stand silo-1
1176858039 daily-inc /stand silo-1
1175908552 full /stand silo-1
The time stamps listed match the time the backup started to write. I
don't keep logs long enough to verify the cumulative and full backup
times, but the daily-inc backups are correct. With this raw data, I can
start parsing and formatting in a number of ways.
I have been blessed with a number of session failures due to a faulty
SAN switch, 1 known bad drive, and probably 1 more once I start
rummaging through logs :-) This has allowed me to see how this reacts
to failed backups. The failed backups do not show even though they ran
after completed backups. YMMV.
Jeff
Darren Dunham wrote:
>>I know things like this have been discussed but I haven't found a way to
>>get what I want from the command line. It looks like I need a little
>>bit of output from bpdbjobs -all_columns, using the lastbackup column
>>definition in the bp.conf file, and part of the bpcatlist command. I
>>have spent considerable time looking on the web and RTFM (several of
>>them). Here is what I'm trying to get. If you know of a command that
>>gives all of this or have a script that will dig all this up, I would
>>greatly appreciate it.
>>
>>Client name, file system or mount point, last successful full date, last
>>successful incremental date. I want something like:
>>
>>client1,/,04/01/2007 19:20:00, 04/16/2007 20:00:03
>>client1,/mnt/data,04/01/2007 19:38:00, 04/16/2007 21:33:00
>>
>>
>
>I don't have anything that does exactly that, but it's close. I can at
>least tell you how I'm getting the data:
>
>Use 'bppllist' to get the list of clients in active policies.
>
>For each of those clients, get a list of images in the time frame of
>interest (can limit to full at this stage if you want) with 'bpimagelist'.
>
>For each of those images, use 'bpflist' to get the name of the first
>file in the backup. That's normally the filesystem or directory that is
>being backed up.
>
>Sort the ones with the same name by date and report the last one.
>
>'bpdbjobs' would give you the filesystem mapping a bit more directly,
>but then you're limited to what's in the jobs list. If you kept that
>and stored it in your own database, that would be easier to manipulate.
>I went to the file data because it hangs around longer.
>
>
>
>>I'm guessing this is why Aptare and Bocada are in business :-)
>>
>>
>
>Seems to be. :-)
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mailman.eng.auburn.edu/pipermail/veritas-bu/attachments/20070417/32e555dd/attachment.htm
|
